Data, the source of endless knowledge we seek,
A digital treasure trove that grows each day.
With every byte, our understanding grows weak
And yet, we find it more and more to say.
It flows like streams and rivers, vast and wide
From countless sources, both far and near in sight
It brings the world together, side by side
With insights, facts, and knowledge brought to light.
Yet, it’s not just numbers, charts, and code
It’s memories, emotions, and human dreams
It’s a symbol of what’s been, what’s to hold
And what’s to come, it’s the fabric of our seams.
Data, our modern muse, with endless rhyme
A sonnet to the digital age in time.
It has been said if you had infinite monkeys on typewriters, you would eventually end up with the works of Shakespeare – but this took less than twenty seconds to create from the moment I hit the enter key. That’s a lot of monkeys working very quickly, and I am just one lowly monkey with one typewriter that needs frequent tea breaks. It is impressive to say the least.
Aside from the fun (albeit unnerving) speed in which ChatGPT can create poetry, what can it do for data offices and real-world business objectives? It is easy to see how marketing teams could utilise the tool to create new advertisements or how consultants could craft presentations on specific projects, but the uses of this data-intensive tool for the betterment of the data world do seem less obvious.
According to The Times: “Technology innovations are often hard to get a handle on as an executive. GPT is different. It democratises what have been opaque technical processes. For example, it can be used for data analysis by generating natural-language summaries of large data sets, making it easier for non-technical team members to understand and make decisions based on the data.”
The impact that this can have on data literacy initiatives and programmes could potentially be huge. We have heard time and again from DataIQ members how data literacy is a core focus and a large hurdle for different businesses, now, perhaps, there is a quick and easy way to make data literacy accessible on a mass scale. Instructional documents about a specific type of data literacy could be produced in a way that uses ChatGPT to rapidly, and succinctly, summarise data groups that previously required expert examination.
The capabilities of ChatGPT are not necessarily new. Large tech companies like Google and Meta developed their own language model tools using human prompts to devise sophisticated responses in the past. The difference here is that OpenAI, the creators of ChatGPT, created a user interface that allows the public to experiment with – and teach – the software directly. There is no reason why it could not learn from data specialists and evolve into something that is specifically designed for data literacy uses. This is the tip of the iceberg of what will be possible as AI and machine learning continue to reach new heights every month.
Ethics
The downside to all this, as we have seen frequently in the past, is that businesses that rapidly implement new AI solutions often fall flat when it comes to ethics, diversity, bias and data protection. AI speech software has recently been used to imitate persons of interest and make it look like they are saying abhorrent things, and I am sure people remember the problems that first arose with racial bias with facial recognition technology. There is still a long way to go to fix these problems, if they can ever truly be fixed.
There is a disclaimer on the website that ChatGPT:
- May occasionally generate incorrect information.
- May occasionally produce harmful instructions or biased content.
- Limited knowledge of world and events after 2021.
The impacts of this technology also affect all ages of people, not just professionals. There have been fears that it can be used by students to write their homework for them – fortunately, new software that can detect AI-generated text will be available, so teachers and professors can rest easy.
It does need pointing out that ChatGPT is not as advanced as it will eventually be – with some reports that it will be 500 times more powerful in just a few months. It can still create nonsensical and biased responses, which is where the more traditional issues with AI creativity show themselves. As impressive as the text generation can be, it does not have the ability to analyse and evaluate what it creates within the social or ethical context of the world. It needs editing and it needs to be scrutinised, just like any business proposal or summary document – it needs a human to examine its purpose.
As of right now, ChatGPT is not a replacement for, say, an actual human writing an article about poetry and the potential ethics of text creation for data usage (please let me keep my job, DataIQ!).
—
Have you been using ChatGPT or looking into its potential data uses for your organisation? If so, contact DataIQ as we want to hear about your data story.